Cardiac fibroblast miR‐27a may function as an endogenous anti‐fibrotic by negatively regulating Early Growth Response Protein 3 (EGR3)

Abstract Pathological myocardial fibrosis and hypertrophy occur due to chronic cardiac stress. The microRNA‐27a (miR‐27a) regulates collagen production across diverse cell types and organs to inhibit fibrosis and could constitute an important therapeutic avenue. However, its impact on hypertrophy and cardiac remodelling is less well‐known. We employed a transverse aortic constriction (TAC) murine model of left ventricular pressure overload to investigate the in vivo effects of genetic miR‐27a knockout, antisense inhibition of miR‐27a‐5p and fibroblast‐specific miR‐27a knockdown or overexpression. In silico Venn analysis and reporter assays were used to identify miR‐27a‐5p's targeting of Early Growth Response Protein 3 (Egr3). We evaluated the effects of miR‐27a‐5p and Egr3 upon transforming growth factor‐beta (Tgf‐β) signalling and secretome of cardiac fibroblasts in vitro. miR‐27a‐5p attenuated TAC‐induced cardiac fibrosis and myofibroblast activation in vivo, without a discernible effect on cardiac myocytes. Molecularly, miR‐27a‐5p inhibited transforming growth factor‐beta (Tgf‐β) signalling and pro‐fibrotic protein secretion in cardiac fibroblasts in vitro through suppressing the pro‐fibrotic transcription factor Early Growth Response Protein 3 (Egr3). This body of work suggests that cardiac fibroblast miR‐27a may function as an endogenous anti‐fibrotic by negatively regulating Egr3 expression.


| INTRODUC TI ON
microRNAs (miRNAs) are small non-coding RNAs (ncRNAs) approximately 22 nucleotides in length that control expression of genes at the level of transcription, by binding to messenger RNAs (mRNAs). 1 It is becoming increasingly evident that a significant fraction of genes and biochemical pathways are regulated by miRNA or ncRNAs. [2][3][4] Consequently, and unsurprisingly, miRNAs serve numerous functions under healthy and pathological conditions. 5 Involvement of miRNAs in regulation of the cardiovascular system is well-established and includes processes such as chronic stress-induced cardiac remodelling due to aortic stenosis, which is characterized by fibrosis and hypertrophy. 6 Among miRNAs known to regulate cardiac fibrosis are miR-21, miR-29, miR-30 and miR-133, while cardiac hypertrophy is regulated by are miR-212/132, miR-133 and miR-208. [7][8][9][10][11][12][13][14] Fibrosis and hypertrophy are interrelated and can elicit one another, 15 and this is borne out by overlapping miRNAs, such as miR-133, that regulates both these processes.
One notable miRNA-miR-27a-is dysregulated in both animal models of fibrosis and human fibrotic disease. [16][17][18][19] Interestingly, miR-27a has been shown to suppress fibrosis in kidney, bladder, liver and lung pathologies. [16][17][18][19] These findings suggest that targeting miR-27a could constitute a possible method to prevent cardiac fibrosis. However, little is known about miR-27a's role (if any) in cardiac fibrosis. Therefore, better elucidating the net effect of miR-27a on cardiac remodelling will have important ramifications for its potential as a drug target.
To address this gap in our knowledge, we sought to delineate the involvement of miR-27a on cardiac remodelling by using both in vivo and in vitro murine models. We demonstrate that genetically or pharmacologically blocking miR-27a-5p enhanced myocardial fibrosis in vivo, without a discernible effect on cardiomyocytes (CMs).
We also performed in vitro mechanistic studies in cardiac fibroblasts (CFs) to characterize miR-27a-5p function more fully and discovered it inhibited the pro-fibrotic transforming growth factor-beta (Tgf-β) signalling pathway through suppressing the pro-fibrotic transcription factor Early Growth Response Protein 3 (Egr3). This effect resulted in lower CF release of pro-fibrotic proteins. This body of work highlights the beneficial role of CF miR-27a-5p in cardiac remodelling.

| MATERIAL S AND ME THODS
All experiments received approval from the Ethics Review Committee at Hainan General Hospital (Haikou, China). All mice used in this study were male and were housed and cared for according to the guidelines outlined in the National Institutes of Health's (NIH) 'Guide for the Care and Use of Laboratory Animals' (8th edition). The methods are fully detailed in the Supplementary Information.

| Genetic knockout of miR-27a promotes transverse aortic constriction-induced heart fibrosis in vivo
Since we found that miR-27a mimics did not affect CM cell size in vitro ( Figure S1A,B), we next sought to determine whether manipulation of miR-27a levels in vivo may affect heart function in models of cardiac disease. For this purpose, we employed mice with genetic knockout (KO) of miR-27a (miR-27a −/− ). As expected, analysing miR-27a-5p expression in the four major cardiac cell types revealed miR-27a-5p knockdown across the four major cardiac cell types (CFs,  20 ) in miR-27a −/− mice ( Figure 1A). This decrease in miR-27a-5p did not affect baseline cardiac characteristics ( Figure 1B-E) or overall body weight ( Figure S2A).
Next, we assessed whether diminished miR-27a-5p levels would affect cardiac characteristics under pressure overload conditions. Transverse aortic constriction (TAC) procedures were performed to generate a mouse model of pressure overload in the left ventricle. miR-27a −/− animals that underwent TAC displayed no discernable differences in fibrosis levels in lung, liver and kidney tissues from WT TAC mice ( Figure S2B). Notably, miR-27a −/− TAC mice had inferior left ventricular function than WT TAC mice ( Figure 1B). miR-27a −/− TAC mice exhibited greater levels of heart mass ( Figure 1C), left ventricle fibrosis ( Figure 1E

| Administration of miR-27a-5p inhibitor promotes transverse aortic constriction-induced heart fibrosis in vivo
It is possible that compensating mechanisms in miR-27a −/− animals might conceal other effects. Therefore, we evaluated the effect of acute locked nucleic acid (LNA)-based miR-27a-5p inhibition in adult mice by injecting three doses of an anti-miR-27a-5p LNA that targets mmu-miR-27a-5p ( Figure 2A). Anti-miR-27a-5p injections induced a profound decrease in miR-27a-5p levels across the four major cardiac cell types (CFs, CMs, CECs and CVSMCs) in comparison to anti-miR-Ctrl injections at both baseline and TAC-induced conditions ( Figure 2B). As we had observed in mice with genetic KO of miR-27a, anti-miR-27a-5p administration reduced left ventricular function in TAC mice ( Figure 2C).

| Cardiac fibroblast miR-27a-5p levels decline with age and transverse aortic constrictioninduced stress
As modulating miR-27a-5p expression had an effect on cardiac fibrosis without impacting CM size or counts, we undertook a more in-depth examination of CF miR-27a-5p expression in normal and diseased cardiac tissue. Left ventricle-derived CF miR-27a-5p levels declined with time as mice grew older ( Figure S4A), in accordance with its putative role in the regulation of body growth. 21 Notably, CF miR-27a-5p levels dynamically changed following the TAC procedure, with an initial decline 2 days post-surgery and increases thereafter ( Figure S4B). Moreover, miR-27a-5p levels in neonatal rat CFs (NRCFs) and adult mouse CFs were much lower after 2 weeks under continuous culture ( Figure S4C,D). Cumulatively, these results show that cardiac fibroblast miR-27a-5p levels decline with age and TACinduced stress, motivating a more in-depth study on the function of miR-27a-5p in CFs.

| miR-27a-5p suppresses cardiac fibroblast's pro-fibrotic activity via Early Growth Response Protein 3 (Egr3) in vitro
Our in vivo experiments showed that miR-27a-5p displays an antifibrotic effect in TAC mice. We have been suggested that miR-27a-5p may modulate Tgf-β signalling activity in CFs. Indeed, miR-27a-5p mimic in NRCFs decreased bioluminescence in a SBE reporter assay  Figure 5B). We also evaluated the secretome of NRCFs to determine whether miR-27a-5p was associated with secretion of fibrosis-related proteins from NRCFs. Following transfection of NRCFs with anti-miR-27a-5p LNA, the conditioned media were harvested and subjected to proteomic analysis by tryptic digest followed by mass spectrometry ( Figure 5C). Anti-miR-27a-5p LNA in NRCFs increased the secretion of multiple mediators of fibrosis compared to NRCFs transfected with anti-miR-Ctrl LNA ( Figure 5D). Many of the genes encoding for the secreted factors have been associated with the Tgf-β signalling cascade, such as lysyl oxidase-like 3 (Loxl3) 23 and the latent Tgf-β binding proteins (Ltbp1, Ltbp2). 24 Using immunofluorescence, we confirmed that anti-miR-27a-5p LNA enhanced Ltbp1 secretion from NRCFs ( Figure 5E). These combined observations imply that miR-27a-5p negatively regulates pro-fibrotic activity in cardiac fibroblasts.
We then conducted an in silico Venn analysis to determine the target gene(s) of miR-27a-5p that may be responsible for miR-27a-5p's anti-fibrotic action in CFs. We analysed the overlapping putative TargetScan-derived target genes for human, mouse and rat miR-27a-5p that were also up-regulated in the left ventricular transcriptome of WT TAC mice relative to WT sham mice (GEO accession: GSE18224 25 ). This Venn analysis uncovered gasdermin (Gsdma) and Egr3 as potential targets of miR-27a-5p that are up-regulated by TAC ( Figure 5F). Considering that Gsdma is an apoptosis mediator primarily expressed in epithelial cells and T-lymphocytes 26 while Egr3 is a transcription factor involved in the Tgf-β-induced fibrotic response in fibroblasts, 27 we selected Egr3 for further analysis. Follow-up TargetScan analysis revealed two putative miR-27a-5p binding sites in the 3′-untranslated region (3′-UTR) of Egr3 that are conserved across human, mice and rats ( Figure 5G).

| D ISCUSS I ON
Herein, we performed a series of experiments that cumulatively support miR-27a-5p's suppression of pathological cardiac fibrosis during heart remodelling. Selective genetic KO of miR-27a or miR-27a-5p LNA-based inhibition enhanced cardiac stress-induced fibrosis in mice that had undergone TAC procedure without impacting CMs, whereas selective OE of CF miR-27a-5p produced the reverse outcome. Secretome analysis in CFs identified several pro-fibrotic factors that were differentially expressed upon miR-27a KD. Luminescence reporter assays of Egr3 3′-UTR binding pointed to the pro-fibrotic transcription factor Egr3 as a major mediator of miR-27a-5p's suppressive effect on fibrosis.
Several pathological illnesses stimulate fibrosis and the release of extracellular matrix (ECM) proteins. miR-27a has been shown to suppress fibrosis under pathological conditions in the kidney, bladder, and liver in vivo [16][17][18] and CF collagen gene expression in vitro. 12 Therefore, here we have been suggested that miR-27a-5p may have Our 3′-UTR reporter and immunoblotting assays identified Egr3 as a direct regulatory target of miR-27a-5p. We also found that miR-27a-5p KD in NRCFs in vitro promotes pro-fibrotic factor secretion in an Egr3-dependent manner. Egr3 is a member of the Early Growth Response (Egr) family of transcription factors (Egr-1, Egr-2 and Egr-4) that all share a conserved zinc-finger domain targeting the Egr response element present in several gene promoters. 28 Specifically, Egr3 has been shown to be a TGF-ß-induced transcription factor that bolsters pro-fibrotic gene expression in human fibroblasts. 27 Moreover, murine fibroblasts with Egr3 knockout show down-regulated levels of several key fibrotic genes (ie Col1a1, Acta2, Tgfß1, Ctgf and Pai1) in response to Tgf-ß2 stimulus, revealing that Egr3 is necessary and sufficient for Tgf-β-induced fibrotic responses. 27 This is notable considering the Tgf-β2 up-regulation observed in our TAC mouse left ventricular tissue samples. Our proposition is that miR-27a-5p suppresses Tgf-β-induced cardiac fibrosis by inhibiting Egr3 expression in CFs.
Additional research could shed light on several aspects not yet investigated. First, mice can be followed for longer durations after TAC surgery to see if, and to what extent, miR-27a-5p OE decreases cardiac fibrosis over time. Second, researchers can further elucidate the molecular details of miR-27a-5p's mechanism of action to determine how the miR-27a-5p/Egr3 axis affects CFmediated myocardial fibrosis and the significance of the Tgf-β cascade in this biological process. Third, it remains unknown whether miR-27a-5p adopts different distributions in other organs, and whether it shares any overlap in anti-fibrotic function in these organs. We partially addressed this question here, as the impact of miR-27a-5p KO on organ fibrosis was examined in lung, liver and kidney tissues in miR-27a −/− animals following TAC. Histological analysis of these organs did not reflect any ascertainable changes in fibrosis. However, miR-27a-5p's role in fibrotic disease models of these organs remains to be elucidated.
Cumulatively, our results underscore the beneficial role of miR-27a-5p in cardiac remodelling. Other animal models, such as the left-anterior descending coronary artery myocardial infarction (LAD MI) model and Ang infusion model, are needed to confirm the therapeutic potential of miR-27a-5p agomiR therapy in vivo.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest. Data curation (equal); resources (equal); validation (equal).

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.